/*
* Copyright 2012
* Ubiquitous Knowledge Processing (UKP) Lab
* Technische Universität Darmstadt
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package de.tudarmstadt.ukp.dkpro.core.testing.harness;
import static de.tudarmstadt.ukp.dkpro.core.testing.AssertAnnotations.assertSentence;
import static de.tudarmstadt.ukp.dkpro.core.testing.AssertAnnotations.assertToken;
import static org.apache.uima.fit.factory.AnalysisEngineFactory.createEngine;
import static org.apache.uima.fit.util.JCasUtil.select;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
import org.apache.commons.lang.ArrayUtils;
import org.apache.uima.analysis_engine.AnalysisEngine;
import org.apache.uima.analysis_engine.AnalysisEngineDescription;
import org.apache.uima.fit.factory.JCasFactory;
import org.apache.uima.jcas.JCas;
import org.junit.Assert;
import org.junit.internal.AssumptionViolatedException;
import de.tudarmstadt.ukp.dkpro.core.api.resources.ResourceObjectProviderBase;
import de.tudarmstadt.ukp.dkpro.core.api.segmentation.SegmenterBase;
import de.tudarmstadt.ukp.dkpro.core.api.segmentation.type.Paragraph;
import de.tudarmstadt.ukp.dkpro.core.api.segmentation.type.Sentence;
import de.tudarmstadt.ukp.dkpro.core.api.segmentation.type.Token;
import de.tudarmstadt.ukp.dkpro.core.testing.AssertAnnotations;
public final class SegmenterHarness
{
public static final TestData[] DATA = new TestData[] {
new TestData("de.1", "de", "Herr Frank M. Meier hat einen Hund.",
new String[] { "Herr", "Frank", "M.", "Meier", "hat", "einen",
"Hund", "."},
new String[] { "Herr Frank M. Meier hat einen Hund." }),
new TestData("de.2", "de", "Ich bin ein blöder Hund.",
new String[] { "Ich", "bin", "ein", "blöder", "Hund", "." },
new String[] { "Ich bin ein blöder Hund." }),
new TestData("de.3", "de", "Mein Name ist Hans.",
new String[] { "Mein", "Name", "ist", "Hans", "." },
new String[] { "Mein Name ist Hans." }),
// DKPRO-CORE-ASL-98: BreakIteratorSegmenter turns hypens to separate tokens
new TestData("de.4", "de", "ihre Negativbei- spiele immer",
new String[] { "ihre", "Negativbei-", "spiele", "immer" },
new String[] { "ihre Negativbei- spiele immer" }),
new TestData("en.1", "en", "Sadler, A.L. Cha-No-Yu: The Japanese Tea Ceremony.",
new String[] { "Sadler", ",", "A.L.", "Cha-No-Yu", ":", "The",
"Japanese", "Tea", "Ceremony", "."},
new String[] { "Sadler, A.L. Cha-No-Yu: The Japanese Tea Ceremony." } ),
new TestData("en.2", "en", "I love the UIMA toolkit. 1989 is the year in which the Berlin wall fell.",
new String[] { "I", "love", "the", "UIMA", "toolkit", ".",
"1989", "is", "the", "year", "in", "which", "the", "Berlin",
"wall", "fell", "." },
new String[] { "I love the UIMA toolkit.",
"1989 is the year in which the Berlin wall fell." }),
new TestData("en.3", "en", "I'm not a girl.",
new String[] { "I", "'m", "not", "a", "girl", "." },
new String[] { "I'm not a girl." }),
new TestData("en.4", "en", "I am a stupid dog.",
new String[] { "I", "am", "a", "stupid", "dog", "." },
new String[] { "I am a stupid dog." }),
new TestData("en.5", "en", "Georg \"Bullseye\" Logal is a though guy.",
new String[] { "Georg", "\"", "Bullseye", "\"", "Logal",
"is", "a", "though", "guy", "." },
new String[] { "Georg \"Bullseye\" Logal is a though guy." }),
new TestData("en.6", "en", "This doesn't compute.",
new String[] { "This", "does", "n't", "compute", "." },
new String[] { "This doesn't compute." }),
new TestData("en.7", "en", "based on\n 'Carnival of Souls', written by [...] and directed by [...].",
new String[] { "based", "on", "'", "Carnival", "of", "Souls",
"'", ",", "written", "by", "[", "...", "]", "and", "directed",
"by", "[", "...", "]", "." },
new String[] { "based on\n 'Carnival of Souls', written by [...] and directed by [...]." }),
new TestData("en.8", "en", ", , ,",
new String[] { ",", ",", "," },
new String[] { ", , ," }),
new TestData("en.9", "en", "How to tokenize smileys? This is a good example. >^,,^< :0 3:[",
new String[] { "How", "to", "tokenize", "smileys", "?", "This", "is", "a", "good", "example.", ">^,,^<", ":0", "3:[" },
new String[] { "How to tokenize smileys?", "This is a good example.", ">^,,^< :0 3:[" }),
// Sombody who can read arabic, please check this
// Covering the following sub-Saharan countries with vast areas very
new TestData("ar.1", "ar", "تغطي الصحراء الكبرى الدول التالية بمساحات شاسعة جدا",
new String[] { "تغطي", "الصحراء", "الكبرى", "الدول", "التالية",
"مساحات", "شاسعة", "جدا" },
new String[] { "تغطي الصحراء الكبرى الدول التالية بمساحات شاسعة جدا" }),
// While the stanford parser should come with a proper tokenizer
// for Chinese (because it can parse chinese text), this does not
// seem to be the right one or I am using it wrong. The associated
// test cases do not work. Maybe debugging the command below
// would help to find out how to use it.
// They use command to parse it: java -mx1g -cp "stanford-parser.jar"
// edu.stanford.nlp.parser.lexparser.LexicalizedParser -tLPP
// edu.stanford.nlp.parser.lexparser.ChineseTreebankParserParams -sentences
// newline -escaper
// edu.stanford.nlp.trees.international.pennchinese.ChineseEscaper
// -outputFormat "penn,typedDependencies" -outputFormatOptions
// "removeTopBracket" xinhuaFactoredSegmenting.ser.gz sampleInput.txt.
new TestData("zh.1", "zh", "服务业成为广东经济转型升级的重要引擎。",
new String[] {"服务业", "成为", "广东", "经济", "转型", "升级", "的",
"重要", "引擎", "。"},
new String[] {"服务业成为广东经济转型升级的重要引擎。"}),
new TestData("zh.2", "zh", "中国离世界技术品牌有多远?",
new String[] {"中国", "离", "世界", "技术", "品牌", "有", "多远",
"？" },
new String[] { "中国离世界技术品牌有多远?" })
};
private SegmenterHarness()
{
// No instances
}
@FunctionalInterface
public static interface AssumeResourcePredicate {
void assume(String aLanguage, String aVariant)
throws AssumptionViolatedException, IOException;
}
public static void run(AnalysisEngineDescription aAed, String... aIgnoreIds)
throws Throwable
{
run(aAed, null, aIgnoreIds);
}
public static void run(AnalysisEngineDescription aAed, AssumeResourcePredicate aCheck,
String... aIgnoreIds)
throws Throwable
{
// No automatic downloading from repository during testing. This makes sure we fail if
// models are not properly added as test dependencies.
if (offline) {
System.setProperty(ResourceObjectProviderBase.PROP_REPO_OFFLINE, "true");
}
offline = true;
AnalysisEngine ae = createEngine(aAed);
JCas jCas = ae.newJCas();
List<String> results = new ArrayList<String>();
try {
for (TestData td : DATA) {
System.out.printf("== %s ==%n", td.id);
jCas.reset();
if (aCheck != null) {
try {
aCheck.assume(td.language, null);
}
catch (AssumptionViolatedException e) {
results.add(String.format("%s skipped", td.id));
continue;
}
}
jCas.setDocumentLanguage(td.language);
jCas.setDocumentText(td.text);
boolean failed = false;
try {
ae.process(jCas);
AssertAnnotations.assertSentence(td.sentences, select(jCas, Sentence.class));
AssertAnnotations.assertToken(td.tokens, select(jCas, Token.class));
results.add(String.format("%s OK", td.id));
}
catch (Throwable e) {
failed = true;
if (!ArrayUtils.contains(aIgnoreIds, td.id)) {
results.add(String.format("%s FAIL", td.id));
throw e;
}
else {
results.add(String.format("%s FAIL - Known, ignored", td.id));
}
}
if (!failed && ArrayUtils.contains(aIgnoreIds, td.id)) {
results.add(String.format("%s FAIL", td.id));
Assert.fail(td.id + " passed but was expected to fail");
}
}
}
finally {
System.out.println("=== RESULTS ===");
for (String r : results) {
System.out.println(r);
}
}
}
public static void testZoning(Class<? extends SegmenterBase> aSegmenter)
throws Exception
{
testZoning(aSegmenter, "en");
}
public static void testZoning(Class<? extends SegmenterBase> aSegmenter, String aLanguage)
throws Exception
{
testLaxZoning(aSegmenter, aLanguage);
testStrictZoning(aSegmenter, aLanguage);
testOufOfBoundsZones(aSegmenter, aLanguage);
}
public static void testLaxZoning(Class<? extends SegmenterBase> aSegmenter, String aLanguage)
throws Exception
{
// No automatic downloading from repository during testing. This makes sure we fail if
// models are not properly added as test dependencies.
if (offline) {
System.setProperty(ResourceObjectProviderBase.PROP_REPO_OFFLINE, "true");
}
offline = true;
String[] sentences = { "A a a a .", "A a a a -", "B b b b .", "B b b b -", "C c c c .",
"C c c c -" };
String[] tokens = { "A", "a", "a", "a", ".", "A", "a", "a", "a", "-", "B", "b", "b", "b",
".", "B", "b", "b", "b", "-", "C", "c", "c", "c", ".", "C", "c", "c", "c", "-" };
JCas jcas = JCasFactory.createJCas();
jcas.setDocumentLanguage(aLanguage);
// 1 1 2 2 3 3 4 4 5 5 6
// 0 5 0 5 0 5 0 5 0 5 0 5 0
// ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
jcas.setDocumentText("A a a a . A a a a - B b b b . B b b b - C c c c . C c c c -");
// |------------------| |------------------|
new Paragraph(jcas, 0, 19).addToIndexes();
new Paragraph(jcas, 40, 59).addToIndexes();
AnalysisEngine ae = createEngine(aSegmenter,
SegmenterBase.PARAM_STRICT_ZONING, false,
SegmenterBase.PARAM_ZONE_TYPES, Paragraph.class);
ae.process(jcas);
assertToken(tokens, select(jcas, Token.class));
assertSentence(sentences, select(jcas, Sentence.class));
}
public static void testOufOfBoundsZones(Class<? extends SegmenterBase> aSegmenter,
String aLanguage)
throws Exception
{
// No automatic downloading from repository during testing. This makes sure we fail if
// models are not properly added as test dependencies.
if (offline) {
System.setProperty(ResourceObjectProviderBase.PROP_REPO_OFFLINE, "true");
}
offline = true;
// 1 1 2 2 3 3 4 4 5 5 6
// 0 5 0 5 0 5 0 5 0 5 0 5 0
// ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
String text = "A a a a . A a a a - B b b b . B b b b - C c c c . C c c c -";
// |------------------| |------------------|
// non-strict zoning
{
String[] sentences = { "A a a a .", "A a a a -", "B b b b .", "B b b b -", "C c c c .",
"C c c c -" };
String[] tokens = { "A", "a", "a", "a", ".", "A", "a", "a", "a", "-", "B", "b", "b",
"b", ".", "B", "b", "b", "b", "-", "C", "c", "c", "c", ".", "C", "c", "c", "c",
"-" };
JCas jcas = JCasFactory.createJCas();
jcas.setDocumentLanguage(aLanguage);
jcas.setDocumentText(text);
new Paragraph(jcas, 0, 19).addToIndexes();
new Paragraph(jcas, 40, 65).addToIndexes();
AnalysisEngine ae = createEngine(aSegmenter,
SegmenterBase.PARAM_STRICT_ZONING, false,
SegmenterBase.PARAM_ZONE_TYPES, Paragraph.class);
ae.process(jcas);
assertToken(tokens, select(jcas, Token.class));
assertSentence(sentences, select(jcas, Sentence.class));
}
// strict zoning
{
String[] sentences = { "A a a a .", "A a a a -", "C c c c .", "C c c c -" };
String[] tokens = { "A", "a", "a", "a", ".", "A", "a", "a", "a", "-", "C", "c", "c",
"c", ".", "C", "c", "c", "c", "-" };
JCas jcas = JCasFactory.createJCas();
jcas.setDocumentLanguage(aLanguage);
jcas.setDocumentText(text);
new Paragraph(jcas, 0, 19).addToIndexes();
new Paragraph(jcas, 40, 65).addToIndexes();
AnalysisEngine ae = createEngine(aSegmenter,
SegmenterBase.PARAM_STRICT_ZONING, true,
SegmenterBase.PARAM_ZONE_TYPES, Paragraph.class);
ae.process(jcas);
assertToken(tokens, select(jcas, Token.class));
assertSentence(sentences, select(jcas, Sentence.class));
}
}
public static void testStrictZoning(Class<? extends SegmenterBase> aSegmenter, String aLanguage)
throws Exception
{
// No automatic downloading from repository during testing. This makes sure we fail if
// models are not properly added as test dependencies.
if (offline) {
System.setProperty(ResourceObjectProviderBase.PROP_REPO_OFFLINE, "true");
}
offline = true;
String[] sentences = { "A a a a .", "A a a a -", "C c c c .", "C c c c -" };
String[] tokens = {
"A", "a", "a", "a", ".",
"A", "a", "a", "a", "-",
"C", "c", "c", "c", ".",
"C", "c", "c", "c", "-" };
JCas jcas = JCasFactory.createJCas();
jcas.setDocumentLanguage(aLanguage);
// 1 1 2 2 3 3 4 4 5 5 6
// 0 5 0 5 0 5 0 5 0 5 0 5 0
// ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
jcas.setDocumentText("A a a a . A a a a - B b b b . B b b b - C c c c . C c c c -");
// |------------------| |------------------|
new Paragraph(jcas, 0, 19).addToIndexes();
new Paragraph(jcas, 40, 59).addToIndexes();
AnalysisEngine ae = createEngine(aSegmenter,
SegmenterBase.PARAM_STRICT_ZONING, true,
SegmenterBase.PARAM_ZONE_TYPES, Paragraph.class);
ae.process(jcas);
assertToken(tokens, select(jcas, Token.class));
assertSentence(sentences, select(jcas, Sentence.class));
}
static class TestData
{
final String id;
final String language;
final String text;
final String[] sentences;
final String[] tokens;
public TestData(String aId, String aLanguage, String aText, String[] aTokens, String[] aSentences)
{
id = aId;
language = aLanguage;
text = aText;
sentences = aSentences;
tokens = aTokens;
}
}
private static boolean offline = true;
public static void autoloadModelsOnNextTestRun()
{
offline = false;
}
}